Bump version to 1.3.0

README.md

@@ -23,7 +23,7 @@ NNS is built around partial moments, the lower and upper components of variance,
 |---|---|
 | Distribution package | `ovvo-nns` |
 | Import package | `nns` |
-| Current version | `1.2.0` |
+| Current version | `1.3.0` |
 | Python | `>=3.11` |
 | Required runtime dependencies | NumPy, SciPy |
 | R required at runtime | No |

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "ovvo-nns"
-version = "1.2.0"
+version = "1.3.0"
 description = "Python port of nonlinear nonparametric statistics from R NNS"
 readme = "README.md"
 requires-python = ">=3.11"

src/nns/__init__.py

@@ -4,7 +4,7 @@
 
 from nns.pm_matrix import pm_matrix as pm_matrix
 
-__version__ = "1.2.0"
+__version__ = "1.3.0"
 
 _EXPORTS = {
     "BoostResult": ("nns.boost", "BoostResult"),

tests/benchmarks/_r_baseline.json

@@ -78,6 +78,6 @@
     "pm_matrix_50x500_seconds": 0.0072,
     "sd_efficient_set_50x252_degree2_seconds": 0.0044
   },
-  "nns_version": "12.1",
+  "nns_version": "13.0",
   "schema_version": 1
 }

tests/benchmarks/test_lpm.py

@@ -896,7 +896,6 @@ def test_nns_boost_factor_predictor_50x2(
         variable[:10],
         learner_trials=10,
         cv_size=0.25,
-        factor_levels=(["low", "mid", "high"], None),
         feature_importance=False,
     )
 
@@ -926,7 +925,6 @@ def test_nns_boost_multi_factor_predictor_50x3(
         variable[:10],
         learner_trials=10,
         cv_size=0.25,
-        factor_levels=(["low", "mid", "high"], None, ["down", "up"]),
         feature_importance=False,
         random_seed=1,
     )

tests/invariants/test_stack.py

@@ -36,6 +36,21 @@ def test_nns_stack_numeric_shapes_and_keys() -> None:
     assert np.all(np.isfinite(result["stack"]))
 
 
+def test_nns_stack_accepts_duplicate_predictor_columns() -> None:
+    # Parity with nns_reg: the cbind(x, x) dimension trick must work for
+    # nns_stack too. R NNS.stack historically rejected duplicate predictor
+    # names ("[IVs.train] predictor names must be unique."); the Python port
+    # never had that restriction and must keep accepting duplicate columns in
+    # both the training matrix and the point-estimate matrix.
+    rng = np.random.default_rng(123)
+    x = rng.standard_normal(60)
+    y = rng.standard_normal(60)
+    variable = np.column_stack((x, x))
+
+    result = nns_stack(variable, y, variable[:5], method=(1,), folds=2, cv_size=0.25)
+
+    assert result["reg"].shape == (5,)
+    assert np.all(np.isfinite(result["reg"]))
 
 
 def test_nns_stack_pred_int_falls_back_to_point_estimate_when_regression_drops_rows() -> None: